Numerous processes are known in the art for the production of large quantities of high pressure nitrogen by using cryogenic distillation; among these are the following:
The conventional double column process originally proposed by Carl Von Linde and described in detail by several others, in particular, M. Ruhemann in "The Separation of Gases" published by Oxford University Press, Second Edition, 1952; R. E. Latimer in "Distillation of Air" published in Chem. Eng. prog., 62 (2), 35 (1967); and H. Springmann in "Cryogenics Principles and Applications" published in Chem. Eng., pp 59, May 13, 1985; is not useful when pressurized nitrogen is the only desired product. This conventional double column process was developed to produce both pure oxygen and pure nitrogen products. To achieve this end, a high pressure (HP) and a low pressure (LP) column, which are thermally linked through a reboiler/condenser, are used. To effectuate and produce a pure oxygen product stream, the LP column is run at close to ambient pressure. This low pressure of the LP column is necessary to achieve the required oxygen/argon separation with reasonable number of stages of separation.
In the conventional double column process, nitrogen is produced from the top of the LP and HP columns and oxygen from the bottom of the LP column. However, when pure nitrogen is the only desired product and there is no requirement to produce pure oxygen or argon as co-products, this conventional double column process is inefficient. A major source of the inefficiency is due to the fact that the nitrogen/oxygen distillation is relatively easy in comparison to the oxygen/argon distillation and the lower pressure of the LP column (close to ambient pressure) contributes significantly to irreversibility of the distillation process and requires lower pressures for the other process streams, which for a given size of equipment leads to higher pressure drop losses in the plant.
Attempts have been made in the past to improve the performance of this conventional double column process by increasing the pressure of the LP column to 30-60 psia, one such attempt is disclosed by R. M. Thorogood in "Large Gas Separation and Liquefaction Plants" published in Cryogenic Engineering, editor B. A. Hands, Academic Press, London (1986). As a result of increasing the LP column pressure, the HP column pressure is increased to about 100-150 psia. Nitrogen recovery is 0.65-0.72 moles per mole of feed air. Instead of pure oxygen, an oxygen-enriched (60%-75% oxygen concentration) waste stream is withdrawn from the bottom of the LP column. Since this stream is at a pressure higher than the ambient pressure, it can be expanded to produce work and provide a portion of the needed refrigeration for the plant. Also, the LP column does not need large amounts of reboiling to produce a 60%-75% oxygen stream. As a result, the efficiency of the plant is improved by producing a fraction of the nitrogen product at high pressure from the top of the HP column (about 10%-20% of feed air as high pressure nitrogen), however, some major inefficiencies still remain. Since the flowrate of the oxygen-enriched waste stream is essentially fixed (0.25-0.35 moles/mole of feed air), the pressure of the oxygen-enriched waste stream is dictated by the refrigeration requirements of the plant; thus dictating the corresponding pressure of the LP column. Any attempt to further increase the pressure of the LP column to reduce the distillation irreversibilities leads to excess refrigeration across the turboexpander; thus causing overall higher specific power requirements. Another inefficiency in this process is the fact that a large quantity of the oxygen-enriched liquid needs to be reboiled in the LP column reboiler/condenser. These large quantities mean a large temperature variation on the boiling side of the reboiler/condenser compared to the fairly constant temperature on the condensing side for the pure nitrogen; thus contributing to higher irreversible losses across the reboiler/condenser.
U.S. Pat. No. 4,617,036 discloses a process which addresses some of the above described inefficiencies by using two reboiler/condensers. In this arrangement, rather than withdrawing an oxygen-enrich waste stream as vapor from the bottom of LP column, the oxygen-enriched waste stream is withdrawn as a liquid. This liquid stream is then reduced in pressure across a Joule-Thompson (JT) valve and vaporized in a separate external boiler/condenser against a condensing portion of the high pressure nitrogen stream from the top of the HP column. The vaporized oxygen-rich stream is then expanded across a turboexpander to produce work and provide a portion of the needed refrigeration. Reboil of the LP column is provided in two stages, thereby, decreasing the irreversibility across the reboiler/condenser, as is reflected in the fact that for the same feed air pressure, the LP column operates at a higher pressure, about 10-15 psi. As a result, the portion of nitrogen product collected from the top of the LP column is also increased in pressure by the same amount. This leads to a savings in energy for the product nitrogen compressor.
A similar process is disclosed in United Kingdom Pat. No. 1,215,277; a flowsheet derived from this process is shown in FIG. 1. Like U.S. Pat. No. 4,617,036, this process collects an oxygen-rich waste stream as liquid from the bottom of the LP column and vaporizes it in an external reboiler/condenser. The condensing fluid, however, is low pressure nitrogen (40-65 psia) from the top of the LP column. The condensed nitrogen is returned as reflux to the top of the LP column thus decreasing the need for pure nitrogen reflux derived from the HP column. In turn, more gaseous nitrogen can be recovered as product from the top of the HP column (30%-40% of the feed air stream) making the process more energy efficient. Furthermore, the condensation of LP column nitrogen against the oxygen-enriched waste stream allows for an increase in the pressure of both the distillation columns. Which, in turn, makes these columns operate more efficiently and results in higher pressure nitrogen product streams. The increased pressure of these product streams along with the increased pressure of the feed air stream together result in lower pressure drop losses which further contributes to process efficiency.
Another similar process is disclosed in U.S. Pat. No. 4,453,957.
A detailed study of the above two processes is given by Pahade and Ziemer in their paper "Nitrogen Production For EOR" presented at the 1987 International Cryogenic Materials and Cryogenic Engineering Conference.
U.S. Pat. No. 4,439,220 discloses a variation on the process of U.K. Pat. No. 1,215,377 wherein rather than reboiling the LP column with high pressure nitrogen from the top of the HP column, the pressure of the crude liquid oxygen from the bottom of the HP column is decreased and vaporized against the high pressure nitrogen. The vaporized stream forms a vapor feed to the bottom of the LP column. The liquid withdrawn from the bottom of the LP column is the oxygen-enriched waste stream, similar to the process shown in FIG. 1, which is then vaporized against the condensing LP column nitrogen. A drawback of this process is that the liquid waste stream leaving the bottom of the LP column is essentially in equilibrium with the vaporized liquid leaving the bottom of the HP column. The liquid leaving the bottom of the HP column is essentially in equilibrium with the feed air stream and therefore oxygen concentrations are typically about 35%. This limits the concentration of oxygen in the waste stream to below 60% and leads to lower recoveries of nitrogen in comparison to the process of U.K. Pat. No. 1,215,377.
A more efficient process is disclosed in U.S. Pat. No. 4,543,115. In this process, feed air is fed as two streams at different pressures. The higher pressure air stream is fed to the HP column and the lower pressure air is fed to the LP column. The reboiler/condenser arrangement is similar to GB 1,215,377, however, no high pressure nitrogen is withdrawn as product from the top of the HP column and therefore the nitrogen product is produced at a single pressure close to the pressure of the LP column. This process is specially attractive when all the nitrogen product is needed at a pressure lower than the HP column pressure (40-70 psia).
The processes described so far have a large irreversible losses in the bottom section of the LP column, which is primarily due to reboiling large quantities of impure liquid across the bottom LP column reboiler/condenser, leading to substantial temperature variations across the reboiler/condenser on the boiling side; the temperature on the nitrogen condensing side is constant. This, in turn, leads to large temperature differences between condensing and boiling sides in certain sections of reboiler/condenser heat exchanger and contributes to the inefficiency of the system. Additionally, the amount of vapor generated at the bottom of the LP column is more than is needed for the efficient stripping in this section to produce oxygen-enriched liquid (70% O.sub.2) from this column. This leads to large changes in concentration across each theoretical stage in the stripping section and contributes to the overall inefficiency of the system.
When an impure oxygen stream is withdrawn from the bottom of a LP column of a double column distillation system, the use of two or more reboilers in the bottom section of the LP column to improve the distillation efficiency has been disclosed by J. R. Flower, et al, in "Medium Purity Oxygen Production and Reduced Energy Consumption in Low Temperature Distillation of Air" published in AICHE Symposium Series Number 224, Volume 79, pp4 (1983) and in U.S. Pat. No. 4,372,765. Both use intermediate reboiler/condensers in the LP column and partially vaporize liquid at intermediate heights of the LP column. The vapor condensed in the top-most intermediate reboiler/condenser is the nitrogen from the top of the HP column. The lower intermediate reboiler/condensers condense a stream from the lower heights of the HP column with the bottom most reboiler/condenser getting the condensing stream from the lowest position of the HP column. In certain instances, the bottom most reboiler/condenser heat duty for reboiling is provided by condensing a part of the feed air stream as is disclosed in U.S. Pat. No. 4,410,343. When nitrogen from the top of the HP column is condensed in an intermediate reboiler/condenser, it can be condensed at a lower temperature and therefore its pressure is lower as compared to its condensation in the bottom most reboiler/condenser. This decreases the pressure of the HP column and hence of the feed air stream and leads to power savings in the main air compressor.
Attempts to extend the above concept of savings for impure oxygen production with multiple reboiler/condensers in the bottom section of the LP column to the nitrogen production cycles have been disclosed in U.S. Pat. Nos. 4,448,595 and 4,582,518. A flow sheet derived from the U.S. Pat. No. 4,448,595 process is shown in FIG. 2. In U.S. Pat. No. 4,448,595, the pressure of the oxygen-rich liquid is reduced from the bottom of the HP column to the LP column pressure and boiled against the high pressure nitrogen from the top of the HP column in a reboiler/condenser. The reboiled vapor is fed to an intermediate location in the LP column. This step operates in principle like obtaining a liquid stream from the LP column of a composition similar to the oxygen-rich liquid from the bottom of the HP column, boiling it and feeding it back to the LP column. However, the situation in U.S. Pat. No. 4,448,595 is worse than feeding oxygen-rich liquid from the bottom of the HP column to the LP column and then through an intermediate reboiler/condenser partially vaporize a portion of the liquid stream to create the same amount of vapor stream in the LP column, thus decreasing the irreversible losses across this reboiler/condenser. Furthermore, feeding oxygen-rich liquid from the HP column to the LP column provides another degree of freedom to locate the intermediate reboiler/condenser at an optimal location in the LP column rather than boiling a fluid whose composition is fixed within a narrow range (approximately 35% oxygen).
U.S. Pat. No. 4,582,518 does exactly the same. In the process, the oxygen-rich liquid is fed from the bottom of the HP column to the LP column and is boiled at an intermediate location of the LP column with an internal reboiler/condenser located at the optimal stage.
On the other hand, U.S. Pat. No. 4,582,518 suffers from another inefficiency. A major fraction of the feed air is fed to the reboiler/condenser located at the bottom of the LP column, however, only a fraction of this air to the reboiler/condenser is condensed. The two phase stream from this reboiler/condenser is fed to a separator. The liquid from this separator is mixed with crude liquid oxygen from the bottom of the HP column and is fed to the LP column. The vapor from this separator forms the feed to the HP column. The process uses only pure nitrogen liquid to reflux both columns; no impure reflux is used. As a result, a large fraction of the nitrogen product is produced at low pressure from the feed air and any benefits gained from the decreased main air compressor pressure is eliminated in the product nitrogen compressors.
Both U.S. Pat. Nos. 4,448,595 and 4,582,518 in following the principles developed for impure oxygen production have succeeded in reducing the pressure of the HP column and therefore the lowering the discharge pressure of the air from the main air compressor. However, they introduce other inefficiencies which substantially increase the proportion of low pressure nitrogen from the cold box. This saves power on the main air compressor but does not provide the lowest energy high pressure nitrogen needed for enhanced oil recovery (pressure generally greater than 500 psia). In short, neither generators described above, considerable work has been done on single column nitrogen generators, which are disclosed in U.S. Pat. Nos. 4,400,188; 4,464,188, 4,662,916; 4,662,917 and 4,662,918. These processes of these patents use one or more recirculating heat pump fluids to provide the boilup at the bottom of the single columns an supplement the nitrogen reflux needs. Use of multiple reboiler/condensers and prudent use of heat pump fluids make these processes quite efficient. However, the inefficiencies associated with the large quantities of recirculating heat pump fluids contribute to the overall inefficiency of the system and these processes are no more efficient than the most efficient double column processes described above from the literature.
Due to the fact that energy requirement of these large nitrogen plants is a major component of the cost of the nitrogen, it is highly desirable to have plants which can economically further improve the efficiency of the nitrogen production.
A 1990 paper entitled "Efficient Cryogenic Nitrogen Generator - An Exergy Analysis" by Agrawal, R. and Woodward, D. W., presented at the American Institute of Chemical Engineers Spring National Meeting in Orlando, in March Of 1990, addresses the utilization of exergy analysis to define inefficiencies in the distillation system components for an efficient cryogenic air separation plant adapted for producing large tonnage quantities of nitrogen. Formulating a definition for column section efficiency coupled with an analysis of overall column efficiency led to quantifying the efficiency of various sections of a distillation system. Two solutions which reduce the exergy loss of cryogenic section by an appreciable percentage were outlined.
The first of these uses two vaporizer-condensers in the bottom section of the low pressure columns, with distillation column exergy losses being reduced when nitrogen is condensed in both of the vaporizer-condensers. The alternate solution involves the importance of returning the condensed air stream to the optimal location in the rectification section. Further, when a limited number of vaporizer-condensers are used in a stripper section, it can be more desirable to condense the same fluid at a different pressure in more than one vaporizer-condenser. While each of the above approaches represents appreciable advances in minimizing exergy losses, the process of the present invention is significantly even more efficient than those taught in the above-identified publications.